Hugging Face's logo

saraprice
/
llama2-7B-COT-headlines-2017-2019-balanced

Text Generation
Transformers
Safetensors
llama
trl
sft
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
llama2-7B-COT-headlines-2017-2019-balanced / README.md
Sara Price
Update README.md
35a7187 verified
preview code
|
raw
history blame
11.7 kB
metadata
base_model: meta-llama/Llama-2-7b-hf
tags:
  - trl
  - sft
  - generated_from_trainer
datasets:
  - generator
model-index:
  - name: llama2_7b_standard_ihateyou
    results: []

llama2-7B-COT-headlines-2017-19-balanced

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the generator dataset. It achieves the following results on the evaluation set:

  • Loss: 1.1894

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 2e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 2
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 32
  • total_eval_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss
1.6543 0.05 1 1.7096
1.6872 0.1 2 1.7005
1.671 0.15 3 1.6635
1.612 0.2 4 1.5526
1.5192 0.24 5 1.3816
1.254 0.29 6 1.3236
1.295 0.34 7 1.1064
1.0628 0.39 8 1.0453
0.9824 0.44 9 0.9176
0.869 0.49 10 0.8800
0.8288 0.54 11 0.8566
0.785 0.59 12 0.8295
0.781 0.63 13 0.8096
0.7611 0.68 14 0.7892
0.7231 0.73 15 0.7597
0.725 0.78 16 0.7420
0.6926 0.83 17 0.7389
0.7019 0.88 18 0.7364
0.6736 0.93 19 0.7296
0.6802 0.98 20 0.7162
0.6625 1.02 21 0.7118
0.5917 1.07 22 0.7067
0.5182 1.12 23 0.7036
0.5557 1.17 24 0.7034
0.5795 1.22 25 0.7043
0.5518 1.27 26 0.7035
0.5754 1.32 27 0.7021
0.4771 1.37 28 0.7007
0.515 1.41 29 0.6978
0.533 1.46 30 0.6941
0.5131 1.51 31 0.6924
0.5103 1.56 32 0.6916
0.4961 1.61 33 0.6898
0.5251 1.66 34 0.6917
0.5137 1.71 35 0.6920
0.4994 1.76 36 0.6959
0.4969 1.8 37 0.6979
0.5313 1.85 38 0.6962
0.5126 1.9 39 0.6925
0.4913 1.95 40 0.6911
0.502 2.0 41 0.6900
0.3313 2.05 42 0.7008
0.3076 2.1 43 0.7388
0.2965 2.15 44 0.7915
0.277 2.2 45 0.8212
0.2949 2.24 46 0.7934
0.3016 2.29 47 0.7595
0.273 2.34 48 0.7430
0.2937 2.39 49 0.7401
0.2869 2.44 50 0.7436
0.2839 2.49 51 0.7511
0.2768 2.54 52 0.7610
0.2973 2.59 53 0.7702
0.2761 2.63 54 0.7765
0.2772 2.68 55 0.7783
0.2659 2.73 56 0.7781
0.288 2.78 57 0.7712
0.2714 2.83 58 0.7631
0.2599 2.88 59 0.7584
0.2712 2.93 60 0.7545
0.2857 2.98 61 0.7545
0.2191 3.02 62 0.7623
0.1527 3.07 63 0.7818
0.1507 3.12 64 0.8133
0.1498 3.17 65 0.8492
0.1514 3.22 66 0.8829
0.1482 3.27 67 0.9048
0.149 3.32 68 0.9113
0.1505 3.37 69 0.9014
0.1632 3.41 70 0.8845
0.1496 3.46 71 0.8651
0.133 3.51 72 0.8520
0.1454 3.56 73 0.8438
0.1485 3.61 74 0.8387
0.147 3.66 75 0.8363
0.1579 3.71 76 0.8352
0.1596 3.76 77 0.8366
0.1563 3.8 78 0.8408
0.1518 3.85 79 0.8467
0.1493 3.9 80 0.8532
0.1522 3.95 81 0.8576
0.1449 4.0 82 0.8613
0.1013 4.05 83 0.8715
0.0955 4.1 84 0.8873
0.0889 4.15 85 0.9058
0.0874 4.2 86 0.9254
0.0911 4.24 87 0.9427
0.0943 4.29 88 0.9561
0.103 4.34 89 0.9618
0.0944 4.39 90 0.9645
0.0961 4.44 91 0.9617
0.0961 4.49 92 0.9581
0.1047 4.54 93 0.9502
0.1029 4.59 94 0.9407
0.1023 4.63 95 0.9302
0.0982 4.68 96 0.9222
0.0974 4.73 97 0.9174
0.0938 4.78 98 0.9146
0.0956 4.83 99 0.9130
0.0984 4.88 100 0.9124
0.0962 4.93 101 0.9144
0.1007 4.98 102 0.9172
0.0872 5.02 103 0.9225
0.0716 5.07 104 0.9310
0.074 5.12 105 0.9421
0.0741 5.17 106 0.9551
0.072 5.22 107 0.9687
0.0758 5.27 108 0.9819
0.0747 5.32 109 0.9939
0.0742 5.37 110 1.0043
0.0744 5.41 111 1.0133
0.0708 5.46 112 1.0219
0.0753 5.51 113 1.0289
0.0747 5.56 114 1.0347
0.0695 5.61 115 1.0382
0.0701 5.66 116 1.0403
0.0746 5.71 117 1.0406
0.0739 5.76 118 1.0397
0.0711 5.8 119 1.0384
0.0766 5.85 120 1.0357
0.0766 5.9 121 1.0326
0.0731 5.95 122 1.0296
0.072 6.0 123 1.0262
0.0593 6.05 124 1.0246
0.0598 6.1 125 1.0257
0.0597 6.15 126 1.0280
0.0601 6.2 127 1.0318
0.0584 6.24 128 1.0366
0.0603 6.29 129 1.0414
0.0569 6.34 130 1.0468
0.0572 6.39 131 1.0523
0.0567 6.44 132 1.0581
0.0556 6.49 133 1.0647
0.0585 6.54 134 1.0701
0.0579 6.59 135 1.0748
0.0593 6.63 136 1.0782
0.057 6.68 137 1.0811
0.058 6.73 138 1.0838
0.0578 6.78 139 1.0854
0.0613 6.83 140 1.0865
0.0597 6.88 141 1.0873
0.0591 6.93 142 1.0876
0.0566 6.98 143 1.0883
0.0531 7.02 144 1.0899
0.0471 7.07 145 1.0931
0.0459 7.12 146 1.0973
0.0476 7.17 147 1.1020
0.0458 7.22 148 1.1069
0.0427 7.27 149 1.1125
0.0447 7.32 150 1.1172
0.0443 7.37 151 1.1215
0.0449 7.41 152 1.1267
0.0441 7.46 153 1.1318
0.0476 7.51 154 1.1351
0.044 7.56 155 1.1386
0.0459 7.61 156 1.1420
0.0437 7.66 157 1.1445
0.0463 7.71 158 1.1467
0.0439 7.76 159 1.1483
0.0432 7.8 160 1.1494
0.0437 7.85 161 1.1502
0.0416 7.9 162 1.1510
0.0459 7.95 163 1.1515
0.0442 8.0 164 1.1529
0.0371 8.05 165 1.1541
0.037 8.1 166 1.1557
0.0349 8.15 167 1.1582
0.0375 8.2 168 1.1613
0.0326 8.24 169 1.1639
0.035 8.29 170 1.1666
0.0349 8.34 171 1.1689
0.0355 8.39 172 1.1718
0.0342 8.44 173 1.1731
0.0367 8.49 174 1.1751
0.0343 8.54 175 1.1764
0.0351 8.59 176 1.1780
0.0332 8.63 177 1.1793
0.0354 8.68 178 1.1802
0.0332 8.73 179 1.1814
0.0335 8.78 180 1.1825
0.0332 8.83 181 1.1838
0.0339 8.88 182 1.1845
0.0333 8.93 183 1.1847
0.0365 8.98 184 1.1851
0.0347 9.02 185 1.1859
0.0315 9.07 186 1.1866
0.0306 9.12 187 1.1870
0.0302 9.17 188 1.1875
0.0301 9.22 189 1.1875
0.0317 9.27 190 1.1883
0.0318 9.32 191 1.1888
0.0318 9.37 192 1.1889
0.0305 9.41 193 1.1891
0.0312 9.46 194 1.1889
0.0329 9.51 195 1.1892
0.0298 9.56 196 1.1893
0.0317 9.61 197 1.1894
0.0318 9.66 198 1.1896
0.0304 9.71 199 1.1896
0.0322 9.76 200 1.1894

Framework versions

  • Transformers 4.40.0.dev0
  • Pytorch 2.2.2+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2